Re: [PATCH Part1 RFC v4 05/36] x86/sev: Define the Linux specific guest termination reasons

From: Borislav Petkov
Date: Tue Aug 10 2021 - 17:51:56 EST


On Tue, Aug 10, 2021 at 02:30:44PM -0500, Tom Lendacky wrote:
> IIRC, during the review of the first GHCB version there was discussion
> about assigning reason sets outside of 0 within the spec and the overall
> feeling was to not do that as part of the spec.
>
> We can re-open that discussion for the next version of the GHCB document.

My worry is that if nothing documents which sets are allocated to which
vendor, it'll become a mess.

Imagine a Linux SNP guest and a windoze one, both running on a KVM
hypervisor (is that even possible?) and both using the same termination
reason set with conflicting reason numbers.

Unneeded confusion.

Unless the spec says, "reason set 1 is allocated to Linux, set 2 to
windoze, etc"

Then all know which is which.

And so on...

Thx.

--
Regards/Gruss,
Boris.

https://people.kernel.org/tglx/notes-about-netiquette